[PATCH v4] kernfs: allocate inodes and dentry outside kernfs_rwsem

[PATCH v4] kernfs: allocate inodes and dentry outside kernfs_rwsem

[Shakeel Butt]

In the Meta fleet, on 6.16 with the kernfs changes from 7.4 backported,
we still see long cgroupfs stalls:

- oomd and below waited 72 seconds in kernfs_fop_readdir() while an out
  of memory container created cgroups.  They resumed when the kernel OOM
  killer killed the container's workload.

- oomd waited 15 minutes in kernfs_iop_lookup().  PID 1 waited for
  kernfs_rwsem in cgroup_rmdir() while holding cgroup_mutex.

- A lock profiler sampling part of the fleet saw kernfs_iop_lookup()
  hold kernfs_rwsem across memcg reclaim on 26 hosts in 29 days, for up
  to 625 ms.

The profiler shows where these stalls come from.  kernfs_iop_lookup()
and the export ops allocate a new inode while holding kernfs_rwsem for
read.  The inode is charged to the caller's memcg.  If that memcg is at
memory.max, the allocation can stay in reclaim for a long time,
sometimes until an OOM kill.  A writer then queues on the lock and every
new reader queues behind it.  On cgroup2 this blocks every cgroupfs user
on the machine.

Such lookups are common in a memcg that is out of memory.  Its reclaim
drops its kernfs dentries fast, and the files of a new cgroup have no
inodes yet.

Allocate the inode before taking kernfs_rwsem, as kernfs_iop_mkdir()
already does, and attach it under the lock with
kernfs_get_inode_locked().  That helper adds a new inode only for an
active node, so __kernfs_remove() still finds every inode it must clear
i_nlink on, as commit dc1a0cc95737 ("kernfs: don't lose IN_DELETE_SELF
when decoding a file handle") needs.

The inode goes unused when the name does not exist, the node is
inactive, or it already has an inode.  In a VM, a lookup of a missing
name takes about 1 us longer, 4.5 us instead of 3.5 us on cgroupfs.
Cold lookups of existing sysfs and cgroupfs files take 2 to 3% longer.

kernfs_get_parent_dentry() also held the lock across d_obtain_alias(),
which allocates a dentry.  Hold it only around the inode lookup, as
__kernfs_fh_to_dentry() already does.  This also gives ->get_parent the
active test that the IN_DELETE_SELF fix left out: a parent that is being
removed now gets -ESTALE.  Before, it got an inode anyway, possibly a
new one whose i_nlink __kernfs_remove() never cleared.

kernfs_fill_super() preallocates the root inode the same way, so no
inode is allocated under kernfs_rwsem any more.

Tested on a 316-CPU machine.  A task in a full 256 MB memcg with slow
swap looks up the files of new cgroups.  Worst waits in 60 seconds:

                                  before    after
  oomd, for kernfs_rwsem          236 ms   0.1 ms
  below, for kernfs_rwsem         356 ms   0.8 ms
  systemd, for kernfs_rwsem       160 ms    37 us
  readdir of another directory    530 ms   3.9 ms

The lookup itself is still slow.  The VFS still holds its directory's
lock, so other work in that one directory still waits.

Suggested-by: Tejun Heo <tj@kernel.org>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Shakeel Butt <shakeel.butt@linux.dev>
---

Christian, this depends on Jori's series in vfs-7.5.directory.  Could you
take it on top of it?  Greg, you asked to wait for -rc1 and for testing
in -next.  Are you OK with this going through the vfs tree?

Changes since v3:
- Rebased on vfs-7.5.directory.  Use the kernfs_get_inode_locked() that
  Jori added there for kernfs_iop_mkdir().
- ->get_parent: a parent that is being removed now gets -ESTALE even if
  it already has an inode, because the helper tests the node first.
- New cost numbers. v3 costs the same as v4 in it.
- Kept Tejun's ack, as the approach he suggested is unchanged.

v3: https://lore.kernel.org/20261016220029.3276154-1-shakeel.butt@linux.dev

Changes since v2:
- Allocate the inode before taking kernfs_rwsem and free it if unused,
  instead of dropping and retaking the lock.  kernfs_iget() becomes
  kernfs_get_inode_locked() (Tejun).
- Do the same in kernfs_fill_super() (Tejun).
- Describe the ->get_parent active test (Tejun).
- Retested on another machine, numbers updated.

v2: https://lore.kernel.org/20261016164345.31618-1-shakeel.butt@linux.dev

Changes since v1:
- dentry allocation outside kernfs_rwsem (Sashiko).
- Couple of comments updated.

v1: https://lore.kernel.org/20261015221110.3183461-1-shakeel.butt@linux.dev

 fs/kernfs/dir.c   | 12 ++++++++++--
 fs/kernfs/inode.c |  3 +++
 fs/kernfs/mount.c | 49 +++++++++++++++++++++++++++++------------------
 3 files changed, 43 insertions(+), 21 deletions(-)

diff --git a/fs/kernfs/dir.c b/fs/kernfs/dir.c
index e72783843edb..fb1dcf35d900 100644
--- a/fs/kernfs/dir.c
+++ b/fs/kernfs/dir.c
@@ -1251,9 +1251,12 @@ static struct dentry *kernfs_iop_lookup(struct inode *dir,
 	struct kernfs_node *parent = dir->i_private;
 	struct kernfs_node *kn;
 	struct kernfs_root *root;
-	struct inode *inode = NULL;
+	struct inode *inode = NULL, *new;
 	const struct ns_common *ns = NULL;
 
+	/* allocate outside kernfs_rwsem, see kernfs_get_inode_locked() */
+	new = new_inode(dir->i_sb);
+
 	root = kernfs_root(parent);
 	down_read(&root->kernfs_rwsem);
 	if (kernfs_ns_enabled(parent))
@@ -1267,9 +1270,13 @@ static struct dentry *kernfs_iop_lookup(struct inode *dir,
 		 */
 		if (!kernfs_active(kn)) {
 			up_read(&root->kernfs_rwsem);
+			iput(new);
 			return NULL;
 		}
-		inode = kernfs_get_inode(dir->i_sb, kn);
+		if (new) {
+			inode = kernfs_get_inode_locked(new, kn);
+			new = NULL;
+		}
 		if (!inode)
 			inode = ERR_PTR(-ENOMEM);
 	}
@@ -1282,6 +1289,7 @@ static struct dentry *kernfs_iop_lookup(struct inode *dir,
 	if (!IS_ERR(inode))
 		kernfs_set_rev(parent, dentry);
 	up_read(&root->kernfs_rwsem);
+	iput(new);
 
 	/* instantiate and hash (possibly negative) dentry */
 	return d_splice_alias(inode, dentry);
diff --git a/fs/kernfs/inode.c b/fs/kernfs/inode.c
index 9cc1a56740b9..53b1a2796e0b 100644
--- a/fs/kernfs/inode.c
+++ b/fs/kernfs/inode.c
@@ -269,6 +269,9 @@ static int kernfs_inode_test(struct inode *inode, void *data)
  * while the inode is inserted and __kernfs_remove() always finds it to
  * clear its i_nlink.  @inode is consumed.
  *
+ * The caller allocates @inode before taking kernfs_rwsem, because the
+ * allocation can block in reclaim and hold up other users of the lock.
+ *
  * Returns the inode now attached to @kn, which is @inode unless one was
  * already hashed for it (e.g. by a file handle decode), or NULL if @kn
  * is no longer active.
diff --git a/fs/kernfs/mount.c b/fs/kernfs/mount.c
index a0be784bfb06..05e8ec1ee7fa 100644
--- a/fs/kernfs/mount.c
+++ b/fs/kernfs/mount.c
@@ -99,8 +99,8 @@ static struct dentry *__kernfs_fh_to_dentry(struct super_block *sb,
 					    int fh_type, bool get_parent)
 {
 	struct kernfs_super_info *info = kernfs_info(sb);
+	struct inode *inode = ERR_PTR(-ESTALE), *new;
 	struct kernfs_node *kn;
-	struct inode *inode;
 	u64 id;
 
 	if (fh_len < 2)
@@ -124,32 +124,32 @@ static struct dentry *__kernfs_fh_to_dentry(struct super_block *sb,
 		return NULL;
 	}
 
+	new = new_inode(sb);
+	if (!new)
+		return ERR_PTR(-ENOMEM);
+
 	/*
-	 * Hold kernfs_rwsem across the lookup as well as kernfs_get_inode().
-	 * __kernfs_remove() deactivates the subtree and clears i_nlink on its
-	 * inodes under the write lock, so under the read lock either
-	 * kernfs_find_and_get_node_by_id() refuses the node, or the inode is
-	 * in the inode hash before the ilookup() pass goes looking for it.
+	 * Hold kernfs_rwsem from the lookup through kernfs_get_inode_locked().
+	 * Then __kernfs_remove() either finds the inode to clear i_nlink, or
+	 * the node is already inactive and no inode is created.
 	 */
 	scoped_guard(rwsem_read, &info->root->kernfs_rwsem) {
 		kn = kernfs_find_and_get_node_by_id(info->root, id);
-		if (!kn)
-			return ERR_PTR(-ESTALE);
-
-		if (get_parent) {
+		if (kn && get_parent) {
 			struct kernfs_node *parent;
 
 			parent = kernfs_get_parent(kn);
 			kernfs_put(kn);
 			kn = parent;
-			if (!kn)
-				return ERR_PTR(-ESTALE);
 		}
-
-		inode = kernfs_get_inode(sb, kn);
-		kernfs_put(kn);
+		if (kn) {
+			inode = kernfs_get_inode_locked(new, kn);
+			new = NULL;
+			kernfs_put(kn);
+		}
 	}
 
+	iput(new);
 	return d_obtain_alias(inode);
 }
 
@@ -171,9 +171,17 @@ static struct dentry *kernfs_get_parent_dentry(struct dentry *child)
 {
 	struct kernfs_node *kn = kernfs_dentry_node(child);
 	struct kernfs_root *root = kernfs_root(kn);
+	struct inode *inode, *new;
 
-	guard(rwsem_read)(&root->kernfs_rwsem);
-	return d_obtain_alias(kernfs_get_inode(child->d_sb, kernfs_parent(kn)));
+	new = new_inode(child->d_sb);
+	if (!new)
+		return ERR_PTR(-ENOMEM);
+
+	scoped_guard(rwsem_read, &root->kernfs_rwsem) {
+		inode = kernfs_get_inode_locked(new, kernfs_parent(kn));
+	}
+
+	return d_obtain_alias(inode);
 }
 
 static const struct export_operations kernfs_export_ops = {
@@ -294,7 +302,7 @@ static int kernfs_fill_super(struct super_block *sb, struct kernfs_fs_context *k
 {
 	struct kernfs_super_info *info = kernfs_info(sb);
 	struct kernfs_root *kf_root = kfc->root;
-	struct inode *inode;
+	struct inode *inode, *new;
 	struct dentry *root;
 
 	info->sb = sb;
@@ -314,8 +322,11 @@ static int kernfs_fill_super(struct super_block *sb, struct kernfs_fs_context *k
 	sb->s_shrink->seeks = 0;
 
 	/* get root inode, initialize and unlock it */
+	new = new_inode(sb);
+	if (!new)
+		return -ENOMEM;
 	down_read(&kf_root->kernfs_rwsem);
-	inode = kernfs_get_inode(sb, info->root->kn);
+	inode = kernfs_get_inode_locked(new, info->root->kn);
 	up_read(&kf_root->kernfs_rwsem);
 	if (!inode) {
 		pr_debug("kernfs: could not get root inode\n");

base-commit: ecdfc51a6a78aaf1c4ac1d8b6ea199cc2ed020ac
-- 
2.53.0-Meta